Path Planning of Land-Air Amphibious Vehicles based on State-Augmented

Land-air amphibious vehicles, innovative mobile robots capable of both aerial flight and terrestrial navigation, have seen significant development with advancements in artificial intelligence. However, the field of amphibious path planning remains underdeveloped. Traditionally, researchers have utilized two distinct algorithms for the separate land and air states, with limited integration in amphibious path planning. Many amphibious vehicles switch states based on path elevation, but integrating mode commands directly into the path nodes information for state switching has proven more efficient. This study introduces a state-classification method based on state augmentation for landair amphibious vehicles. During path planning, we consider the dynamic boundaries and aerodynamic characteristics of flight motion, designing distinct heuristic functions for ground and aerial modes. Utilizing a hybrid A* framework, we develop an amphibious path planning algorithm that generates and optimizes a path accommodating both driving and flying capabilities. Simulation results demonstrate that our vehicles can smoothly transition between states to navigate obstacles, windows, and doors effectively. Benchmark comparisons and real-world experiments confirm the efficiency of our method, showcasing reduced motion switching times and validating the feasibility of rapid maneuvering.